Markus Posted July 2, 2018 Posted July 2, 2018 I just started a character yesterday, banned today. After a total of 3 hours botted, with breaks. Did quests, did mixed combat training, and some magic training. So the play time was human, the progression was human, the mouse clicks and movements were human (right?) and the client isn't detected. So what flaw makes them so damn efficient at spotting bots? I don't get how their bot detection is an industry secret, like someone equally talented to the coders at Jagex should be able to figure out what data points stand out to an anti-bot algorithm. This isn't a rant at Osbot btw, just a rant at botting in general.
Hi G00gle Posted July 2, 2018 Posted July 2, 2018 Might want to look into patterns. People 'always' tend to do the first stuff human-like, but after that they start doing what everyone does: Either fire up the AIO combat (Or whatever script you are using), or splashing. The part after that human-like is seemingly always the same.
Markus Posted July 2, 2018 Author Posted July 2, 2018 1 minute ago, Hi G00gle said: Might want to look into patterns. People 'always' tend to do the first stuff human-like, but after that they start doing what everyone does: Either fire up the AIO combat (Or whatever script you are using), or splashing. The part after that human-like is seemingly always the same. That's a good point, but, then what? Like, when they've recognized that pattern, and start seeing people following it, how do they go about determining that the person is a bot or not? There's gotta be a check or a test, or something definitive, right?
Hi G00gle Posted July 2, 2018 Posted July 2, 2018 I think the key _might_ (Obviously I can not say for sure since I am not a Jagex' employee) lay in the randomizeness. Most scripts always follow the same way on how they progress: They always tend to click on the same things, or have the same behaviour.
Google Posted July 2, 2018 Posted July 2, 2018 (edited) its behavior analysis , when you come out of tutorial island you have so little data to be analyzed so i think it works like a meter, every time it suspects something un-humane happens it adds to your bot percentage basically if your bot percentage goes above 50% you are a bot, so when you are out of tutorial island its so easy to reach 50% because there isn't data available to compare to so your bot meter goes up really high, but when you have an account that has a lot of played time its not easy to get that meter up cause there is a lot of data on that account. This is how fairfight Edited July 2, 2018 by Google 1
Markus Posted July 2, 2018 Author Posted July 2, 2018 2 minutes ago, Google said: its behavior analysis , when you come out of tutorial island you have so little data to be analyzed so i think it works like a meter, every time it suspects something un-humane happens it adds to your bot percentage basically if your bot percentage goes above 50% you are a bot, so when you are out of tutorial island its so easy to reach 50% because there isn't data available to compare to so your bot meter goes up really high, but when you have an account that has a lot of played time its not easy to get that meter up cause there is a lot of data on that account. This is how fairfight That's the thing, with every client so anti-ban focused, shouldn't we have "fixed" all the un-humane things that happen? I feel like every single step of the way is randomized these days. What's missing? Must be something pretty crucial if you reach your hypothetical 50% so quickly, right?
Night Posted July 2, 2018 Posted July 2, 2018 1 minute ago, Markus said: That's the thing, with every client so anti-ban focused, shouldn't we have "fixed" all the un-humane things that happen? I feel like every single step of the way is randomized these days. What's missing? Must be something pretty crucial if you reach your hypothetical 50% so quickly, right? Humans aren't random 1
dragonite3000 Posted July 2, 2018 Posted July 2, 2018 6 minutes ago, Google said: its behavior analysis , when you come out of tutorial island you have so little data to be analyzed so i think it works like a meter, every time it suspects something un-humane happens it adds to your bot percentage basically if your bot percentage goes above 50% you are a bot, so when you are out of tutorial island its so easy to reach 50% because there isn't data available to compare to so your bot meter goes up really high, but when you have an account that has a lot of played time its not easy to get that meter up cause there is a lot of data on that account. This is how fairfight as crazy as this sounds It wouldn't surprise me if part of what you just said was true.
Markus Posted July 2, 2018 Author Posted July 2, 2018 26 minutes ago, Night said: Humans aren't random I guess I just don't get what step of the way you can't humanize. If randomizing isn't the way to go.
RoomScape Posted July 4, 2018 Posted July 4, 2018 On 7/2/2018 at 12:55 PM, Night said: Humans aren't random Computers are more predictable with movements than humans, though. On 7/2/2018 at 12:40 PM, Markus said:
Juggles Posted July 4, 2018 Posted July 4, 2018 There's people bypassing bans. There's ways around it
cammofunk Posted July 4, 2018 Posted July 4, 2018 On 7/2/2018 at 12:55 PM, Night said: Humans aren't random During Prohibition, moonshiners would wear "cow shoes." The fancy footwear left hoofprints instead of footprints, helping distillers and smugglers evade police. That was pretty random of me
Markus Posted July 4, 2018 Author Posted July 4, 2018 15 hours ago, Juggles said: There's people bypassing bans. There's ways around it Why aren't these ways utilized platform wise then? What ways, also?
Juggles Posted July 4, 2018 Posted July 4, 2018 19 minutes ago, Markus said: Why aren't these ways utilized platform wise then? What ways, also? Cause if everyone uses It then it wouldn't work. Trust me I had a way avoid bans for a week botting 24/7 but eventually they caught on and it stopped working